SWE-Bench Verified by OpenAI tests how well a model can solve real bugs in real Python code from GitHub. These bugs are all public information — so the AI models have almost certainly trained on th…
it’s appropriate that you think your brain works like an LLM, because you regurgitated this shitty opinion from somewhere else without giving it any thought at all
Yeah I’m thinking that people who think their brains work like LLM may be somewhat correct. Still wrong in some ways as even their brains learn from several orders of magnitude less data than LLMs do, but close enough.
it’s appropriate that you think your brain works like an LLM, because you regurgitated this shitty opinion from somewhere else without giving it any thought at all
Yeah I’m thinking that people who think their brains work like LLM may be somewhat correct. Still wrong in some ways as even their brains learn from several orders of magnitude less data than LLMs do, but close enough.